Location: Hydrology and Remote Sensing LaboratoryTitle: Remote sensing estimation of evapotranspiration for SWAT Model Calibration Author
|Kustas, William - Bill|
Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 2/16/2012
Publication Date: 5/1/2012
Citation: Sadeghi, A.M., Beeson, P.C., Daughtry, C.S., Akhmedov, B., Alfieri, J.G., Anderson, M.C., Hain, C., Tomer, M.D., Prueger, J.H., Kustas, W.P., Hatfield, J.L. 2012. Remote sensing estimation of evapotranspiration for SWAT Model Calibration [abstract]. ASABE Conference. Paper No. 12-13687. Interpretive Summary:
Technical Abstract: Hydrological models are used to assess many water resource problems from water quantity to water quality issues. The accurate assessment of the water budget, primarily the influence of precipitation and evapotranspiration (ET), is a critical first-step evaluation, which is often overlooked in hydrologic models. Precipitation inputs for hydrologic models are shown to benefit from combining the finer spatial resolution of NEXRAD with the accurate rainfall amounts measured by gauges; however, actual ET resulting from the model is often only assessed as an annual average. Daily measurements of ET values on corn and soybean fields through three growing seasons were used along with MODIS NDVI product to spatially estimate ET throughout the study area for the calibration of Soil and Water Assessment Tool (SWAT) model. The SWAT model offers several possible inputs of potential ET that should be carefully selected to ensure proper estimations of the resulting actual ET and water quantities. For this study, we tested the internally computed potential ET methods using i) Priestley-Taylor, ii) Penman-Monteith, and iii) Hargreaves, in addition to four other sources of potential ET that were read into SWAT. Local and regional weather stations were used to compute two potential ET inputs using Penman-Monteith, along with two large scale products from NCEP/NCAR Reanalysis 1 project and the Atmosphere-Land Exchange Inverse (ALEXI) model. This study was conducted on the South Fork watershed of the Iowa River in central Iowa, U.S.A. The watershed covers about 788 km2 (194,720 ac) and is one of the Benchmark Watersheds of the USDA Conservation Effects Assessment Project (CEAP). Here, reading in potential ET into SWAT computed from local or regional stations performed the best, while the weather generator input only captured the annual average, but not the variation throughout the growing season affecting biomass and yield estimates. If station data is not available, potential ET from the ALEXI model to drive SWAT was better than the weather generator. This research provides insight into the value of remote sensing and field observations for the application of physically-based watershed models, as well as the selection of potential ET inputs to drive the model.